A significant problem in vaccine development is overcoming the effects of the poorly understood phenomenon of immunodominance. Although a particular infectious agent comprises hundreds or thousands of potentially antigenic molecules, each comprising multiple protein or peptide epitopes capable of binding to antibodies or to MHC molecules, the immune response elicited against a particular infectious agent, is often directed against only a limited number of epitopes, or even to a single epitope (see van der Most et al., J. Immunol. 1996 157:5543-54 and references infra). These epitopes are referred to as “immunodominant” epitopes. Such a narrow immune response to a few immunodominant epitopes offers poor protection against subsequent infection by a mutated form or by different strains of the original infectious agent. The problem of immunodominance is especially acute for infectious agents having a high mutation rate and for those comprising multiple strains.
For reasons that are not well understood, the immune response elicited by some immunodominant epitopes actually impairs the ability to develop an effective response against a subsequent infection. This has been observed, for example, with infectious diseases caused by multi-strain pathogens. This phenomenon, also referred to as “original antigenic sin,” was first characterized with respect to the influenza virus (Fazekas de St. Groth and Webster, 1966 J. Exp. Med. 124:331-45), and has since been observed in hepatitis B and C (Harcourt et al., 2003 Clin. Exp. Immunol. 131:122-29), malaria (Good et al., 1993 Parasite Immunol. 15:187-93), dengue fever (Rothman et al., 2001 Vaccine 19, 4694-4699), Chlamydia (Berry et al., 1999 J. Infec. Dis. 179:180-86), and HIV (Anderson et al., 2001 Clin. Immunol. 101:152-57). For example, immunity after infection by a particular strain of dengue virus protects only modestly or even negatively against reinfection by one of the other three strains (Mongkolsapaya et al. 2003 Nature Medicine 9:921-27). This effect has also been observed for subsequent infections by a different pathogen. For example, exposure to influenza appears to increase susceptibility to hepatitis C through immunodominance (Brehm et al. 2002 NI 3:627-34). The existence of this phenomenon means that for some infectious diseases, vaccinated individuals may paradoxically be more susceptible to infection by another strain of the same pathogen, or even to another pathogen, than individuals who were not vaccinated. An effect similar to original antigenic sin has also been observed in the context of tumor immunity (Makki et al. 2002 Cancer Immunity 2:4-17, and references infra; see also Cole et al. 1997 J. Immunol. 158: 4301-09 and van der Most et al. 1996 J. Immunol. 157:554354). Makki et al. suggest that vaccination with certain dominant tumor antigens not only fails to elicit an immune response against the tumor, but also hinders the development of an effective response against other, presumably subdominant tumor antigens.
Immunodominance has also been observed in the context of tumor immunity. For example, an existent response to a tumor antigen may prevent a response to new tumor antigens arising through mutation. Schreiber et al. refers to this phenomenon as the “priority of the first response” which was suggested by experiments in mice showing that repeated immunization with an antigen A, followed by later immunization with an antigen B, fails to elicit an anti-B response (Schreiber et al. 2002 Cancer Biol. 12:25-31, and references infra). This immunodominance could be broken experimentally by vaccination with individual tumor antigens at separate sites, rather than with multiple antigens at one site.
One factor in determining whether an epitope becomes dominant appears to be its effectiveness in generating an immune response (Schreiber et al., Cancer Biol. 2002 12:25-31). This is a function of a number of factors, including the binding affinity of the epitope for T cell receptors or for antibodies expressed by B cells. Intracellular processing of peptide antigens is also a factor because epitopes which are presented at high levels on the surface of antigen presenting cells tend to elicit a stronger response. However, it is not simply the case that the dominant epitopes are able to elicit an immune response, and subdominant epitopes are not. The ability of subdominant epitopes to elicit an immune response has been demonstrated, for example, in the context of viral infections and tumor immunity (see Cole et al, J. Immunol. 1997 158:4301-09; van der Most et al., J. Immunol. 1996 157:5543-54; Makki et al., Cancer Immunity 2002 2:4-17 and references infra). Makki et al. suggests that, in the context of tumor immunity, vaccination with a subdominant epitope may even be superior to vaccination with a dominant epitope.
It is not known why subdominant epitopes which are capable of eliciting an immune response nevertheless often fail to do so. However, there is evidence that dominant epitopes can suppresses immunity to the other, subdominant epitopes. This phenomenon may be a result of the manner in which antigen-specific effector cells are selected. For example, cytotoxic T cells (“CTLs” or “CD8+ T cells”) binding to MHC-peptide complexes on an antigen presenting cell can inhibit the proliferation of other CTLs binding to other complexes on the same cell. Since such binding is required to stimulate T cell proliferation, and only proliferating T cells mature into memory T cells, the effect is presumably to produce narrowing of the repertoire of memory T cells. Thus, the ability to protect against a secondary infection by a similar but not identical infectious agent is reduced. This and other aspects of the cellular biology and immunology of immuodominance in the cytotoxic T cell response against viral infections are reviewed by Yewdell and Del Val, Immunity 2004 2:149-53.
There remain fundamental unanswered questions that have hindered the design of effective vaccines or vaccination strategics that will effectively avoid the adverse effects of immunodominance. For example, the relationship between antibody or T cell receptor binding, the sequence of the antigen, and the emergence of immunodominance is not known. Understanding these relationships is important to vaccine design generally and is of particular importance to the development of safe, effective peptide-based vaccines. For a review of epitope identification, vaccine design and delivery, see Sette and Fikes, 2003 Cur. Opinion Immunol. 15:461-70.
One approach to answering these questions is to utilize mathematical models that capture the sequence-level dynamics of the effector cell and epitope binding interactions. A random energy model is one such mathematical model which has successfully reproduced complex immune phenomena such as immunodominance and original antigenic sin. This model captures much of the thermodynamics of protein folding and ligand binding, and consequently also captures the correlations between the three dimensional amino acid structure of antibodies or T cell receptors and the amino acid sequences of antigenic molecules. The specific antibody or T cell repertoire of an individual is represented in the model by a specific set of amino acid sequences. An epitope of a specific antigen or viral strain is represented by a specific instance of the random parameters. An immune response that finds a T cell receptor or antibody with a high binding affinity to a specific epitope corresponds in the model to finding an amino acid sequence having a low energy for a specific parameter set.
The robustness of the model as a tool for accurately simulating the interactions between effector cells and antigens has been demonstrated by a number of experiments. For example, Deem and Lee demonstrated that original antigenic sin in the context of influenza stems from the localization of the immune system response in antibody sequence space (Phys. Rev. Lett. 2003 91:68101-104). This localization stems from memory sequences being less able to evolve than naïve sequences, and is observed in general for diseases with high year-to-year mutation rates, such as influenza. Building on these results, Deem and Munoz demonstrated that this localization played a role in the ineffectiveness of the 2003-2004 influenza vaccine in the United States (Vaccine 2005 23:1144-48). Predictions from the model also correlated well with the efficacies of the H3N2 influenza A component of the annual vaccine between 1971 and 2004 (Gupta and Deem 2005 Quant. Biol., document no. 0503030). In fact, the predictive value of the model was superior to that of the standard ferret animal model.
In another example, the model was used to examine cross-reactivity in the T cell response to mutated viral antigens (Park and Deem 2004 Physica A. 341:455-70). Here again, the predicted specific lysis curves were in excellent agreement with ex vivo and in vitro altered peptide ligand experiments. Predictions from the model of immunodominance in the human immune response to the four-component dengue vaccine also accurately predicted experimental results (Deem 2004 AIChE J. 50:734-38).
It is clear from these results that the model is able to accurately simulate important aspects of the immune response to an antigen and thereby provide insights for vaccine design and development. The present invention is based in part on such an insight from the model, namely that multi-site vaccination against an infectious agent increases immunity against subdominant epitopes, thereby mitigating the effects of immunodominance.